Ontology-based Validation of Agent Oriented Modelling A. Lopez-Lorca L. Sterling T. Miller G. Beydoun Swinburne University of Technology University of Melbourne University of Wollongong Hawthorn VIC 3122 Australia Parkville VIC 3010 Australia Wollongong NSW 2522 Australia University of Melbourne tmiller@unimelb.edu.au Parkville VIC 3010 Australia {aall645, beydoun}@uow.edu.au lsterling@groupwise.swin.edu.au leon@cs.mu.oz.au ABSTRACT potentially reducing the development and maintenance costs of Despite the potential of Multi-Agent Systems (MAS), this tech- software systems. We provide a methodology-independent and nology has not been widely adopted by industry yet. Due to its ontology-based add-on to facilitate the creation of models espe- complexity, errors in modelling activities can be costly. Early cially in certain scenarios: In developments of inexperienced validation of MAS models can prevent rework or building a sys- modelers, to guide their work and avoid errors; in initial MAS tem non-compliant with client’s specification. We propose a gen- developments of experienced modelers in any other technology, eral ontology-based process to validate any kind of software mod- as agents have many important particularities which cannot be els that can be adapted in a broad range of software development found in other paradigms; in projects where the domain is com- projects. We illustrate this for MAS development as its complex- plex or unknown, for experienced and inexperienced modelers ity justifies additional costs associated with applying our add-on alike; in projects dealing with the same domain, to enable reuse of validation process. This work provides early evidence of the the generated domain knowledge (i.e. ontologies). Whilst the soundness of our approach. We successfully validate and improve focus of our illustrations is on applying ontologies to improve the the quality of MAS models for a real-life development project development of MAS models, we expect that our approach is showing that our MAS models validation process can contribute easily adaptable to other development paradigms such as agile to harnessing the commercial potential of MAS technology. methods. Many existing works focus on the use of ontologies to MAS. Of Categories and Subject Descriptors these many focus on the process itself. For example, in [3], a I.2.11 [Artificial Intelligence]: Distributed Artificial Intelligence method is given to adapt extreme programming methods to de- – Multi Agent Systems; I.6.4 [Simulation and Modeling]: Model velop a lightweight ontology to help agile development of MAS. Validation and Analysis. It is refined further in [4]. Our focus in this paper is the quality of the MAS workproducts through a domain enriched process rather than the software process itself. Other works use ontologies to General Terms assist in the development of workproducts in particular in the Verification. detailed design phase. Tran et al. [5] present an ontology-based MAS for the domain of a peer-to-peer (P2P) information sharing Keywords community where ontologies are built and used in development- Ontology-based, validation, software model, multiagent system time to create the models and in run-time to exchange information model. between agents. Although they use domain ontologies during development and run-time, they do not provide detailed support for the validation of MAS, which is the focus of our proposal. 1. INTRODUCTION Ontologies provide a mechanism for representing domain knowl- Our approach shares similar goals with the work developed by edge to a varying degree of formalism [1]. We advocate the use of Brandão et al. [2]. They propose the use of ontologies as a method ontologies to validate and improve the quality of software work- for the verification of MAS designs. They use an ontology to products during development processes. As an element of joint model the MAS modelling language. These model-diagram map- development with the user, ontologies can bridge common com- pings enable the automatic validation of the models to check that munication gaps between users and developers. We illustrate there are neither intra-model nor inter-model inconsistencies. The using an ontology to check consistency, correctness and com- main difference with our proposal is that they can validate the pleteness of models against initial system requirements. We ex- models against their theoretical structure and dynamics, but use pect that as intermediary modelling elements, ontologies can fa- no information about the specification or application domain. cilitate and improve the development of software workproducts, Furthermore, they do not generalise their efforts to outside MAS development and have not validated their proposal properly. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are 2. AN ONTOLOGY BASED SOFTWARE not made or distributed for profit or commercial advantage and that MODELS VALIDATION ADD-ON copies bear this notice and the full citation on the first page. To copy Our proposed ontology-based MAS software models validation otherwise, or republish, to post on servers or to redistribute to lists, re- quires prior specific permission and/or a fee. (Figure 1) consists of five activities. Our proposal is an add-on to Conference’10, Month 1–2, 2010, City, State, Country. the development process and it is completely independent of the Copyright 2010 ACM 1-58113-000-0/00/0010…$10.00. underlying software models or their development methodology. interaction model were used to prevent errors in the development of the scenario during the second iteration. 3. CONCLUSION AND FUTURE WORK We apply ontologies to validate and improve the quality of soft- ware models. We take into account the domain as specified by the client’s requirements, bridging any communication gap between clients and developers. Models are validated as soon as they are available, fixing errors as they arise and avoiding compounding and propagating errors to later phases of the development. To integrate our validation add-on seamlessly into the development process, we use an iterative, incremental and concurrent devel- opment process. The process iterates over intermediate versions of the model to achieve high quality. It is incremental in nature, not all the models are considered during each iteration. It is con- current as development and validation activities overlap. Applying our process can incur additional development cost and requires a cost justification. It is particularly appealing in critical software application where errors can be very costly and disas- trous. This cost overhead may also be justified in the scenarios described in Section 1. That said, the cost of the validation can be greatly reduced by more effective reuse of existing ontologies. With advent of the Semantic Web, more ontologies are made Figure 1. Ontology-mediated software models validation available. More importantly, there is a great scope for generating add-on process overview. the amendment proposals automatically, harnessing automatic reasoning capabilities of ontologies. Indeed, we are now studying In the Ontology Acquisition activity a suitable ontology is re- this possibility with the expectation to develop a tool that can trieved from an existing repository, otherwise one is built using significantly alleviate the burden of the details of the ontology- the most suitable ontology engineering techniques. In the Ontol- mediated validation process. In the future, we also intend to apply ogy Augmentation activity, the ontology is augmented to repre- the ontology-mediated software model validation process to fur- sent features related to the chosen development paradigm. Do- ther cases studies to fine-tune it and to test our forthcoming tool. main concepts are annotated to link them to paradigm concepts and relations between them are created according to existing rela- 4. REFERENCES tions defined for the paradigm. In our case study we identify the [1] Beydoun, G., A. Hoffmann, J.T. Fernández-Breis, R. MAS terms Goal, Role, Activity, Environment and Agent and Martínez-Béjar, R. Valencia-García and A. Aurum. Coop- relations between them such as Role responsible for Goal, Agent erative Modeling Evaluated. International Journal of Coop- plays Role or Activity follows Activity. In the Ontology Validation erative Information Systems. 2005. 14(1): p. 45-71. activity, members of the development team validate the ontology with the client to reach a common understanding and compliance [2] Brandão, A.A.F., V.T.d. Silva and C.J.P.d. Lucena, Ob- to client’s conceptualisation. In the Software Models Validation served-MAS: An Ontology-Based Method for Analyzing activity, the models are validated against the augmented ontology. Multi-Agent Systems Design Models, in Agent-Oriented This activity provides the control element for new iterations. A Software Engineering VII. 2007. Springer Berlin / Heidel- new iteration will be necessary as long as any non-trivial recom- berg. p. 122-139. mendation is made to improve the quality of the models. In the [3] Hristozova, M. and L. Sterling. An eXtreme Method for De- Software Models Improvement activity, the recommendations are veloping Lightweight Ontologies, in Workshop on Ontologies analysed by the developers to choose which to apply and which to in Agent Systems, at the International Join Conference on ignore. After improving the quality of the models according to Autonomous Agents and Multi-Agent Systems. 2002. Bolo- chosen recommendations, the new set of models will be validated gna, Italy: ACM. in the next iteration. In our case study we validate and improve [4] Lister, K., M. Hristozova and L. Sterling. Reconciling Im- goal, agent, interaction, scenario, organisation, role and environ- plicit and Evolving Ontologies for Semantic Interoperability, ment models for MAS development. in Ontologies for Agents: Theory and Experiences, V. Development proceeds with each iteration further along the se- Tamma, Cranefield, S., Finin, T., Willmott, S, Editor. 2005. quence of workproducts required by the chosen methodology. The Birkhäuser Basel. p. 121-144. development and validation of the software models are inter- [5] Tran, N., G. Beydoun and G. Low. Design of a Peer-to-Peer twined and done concurrently. Problems of reviewed models are Information Sharing MAS Using MOBMAS (Ontology- fixed before their full development. Any models yet to be com- Centric Agent Oriented Methodology), in International Con- menced in that iteration will take advantage of the recommenda- ference on Information Systems Development. 2006. Buda- tions avoiding compounded errors. This has been proven in our pest, Hungary: Springer. case study, as recommendations made in the first iteration to the